Tillage to disrupt (prune) tree roots is an intensive practice which could improve herbage productivity at the crop–tree interface by reducing competition for water. We compared tillage effects on 9- to 11-yr-old loblolly pine (Pinus taeda L.) growth and herbage yields of annual ryegrass (Lolium multiflorum Lam.) and pearl millet [Pennisetum glaucum (L.) R. Br.] on a fragipan soil in Arkansas. Alley crops were rotationally grown in a 9.7-m wide alley (main plot) between bordering trees on one of three tillage treatments: control (surface tillage), rip followed by surface tillage, and trench plus root barrier followed by surface tillage. Topsoil water in May through September, herbage mass, and nutritive value were measured for each crop for 2 or 3 yr in three subplots systematically arrayed (north, middle, and south) across the alley. Diameter at breast height (DBH, measured 1.3 m above soil surface) and height of border trees were measured annually. Trenching resulted in a more uniform distribution of topsoil water among subplots compared to the other tillage treatments. Annual ryegrass yield did not show a tillage response, but pearl millet yielded more herbage in the rip (6760 kg ha–1 in 2003) and trench (3300 kg ha–1 in 2005) than the control treatment (4990 and 1260 kg ha–1 for 2003 and 2005, respectively). Ripping and trenching significantly reduced loblolly pine DBH and height compared to the control. Similarly configured alley cropping practices probably have little potential for annual herbage production even with root pruning.